Perbandingan Algoritma ARIMA, Prophet, dan LSTM dalam Prediksi Penjualan Tiket Wisata Taman Hiburan (Studi Kasus: Saloka Theme Park)

Dienda Rizkya Hayuningtyas Roosaputri (Teknik Informatika, FTI UKSW, Salatiga, Jawa Tengah, Indonesia)
Christine Dewi (Teknik Informatika, FTI UKSW, Salatiga, Jawa Tengah, Indonesia)

Kesatria : Jurnal Penerapan Sistem Informasi (Komputer dan Manajemen)
Vol 4, No 3 (2023): Edisi Juli

Indonesia has a wide variety of tourism destinations or tourism destinations that are useful as a place to have fun with family or friends, to calm down from various activities as well as an educational place for children to be more interested in learning. PT. Panorama Indah Permai (Saloka Theme Park) is one of the tourist destinations in Indonesia and is currently operating again when the Covid-19 pandemic enters Indonesia and results in changes in visitors coming to amusement park tourism. Therefore, a comparison of ARIMA, Prophet, and Long Short-Term Memory (LSTM) algorithms was made to determine the algorithm that is suitable to be used as a predictive model of amusement park ticket sales. The Data obtained comes from historical data and must go through several processes such as pre-processing and algorithm testing to ensure the accuracy of the data with the algorithm. The results obtained from three forecasting models, it was obtained that the ARIMA algorithm has a value of RMSE of 762,009 and MSE of 580,659,053.